ALCPG, Cornell University

1
Scintillator Readout w/ Geiger Avalanche
Photodiodes - update since Arlington

• David Warner, Robert J. Wilson
• Department of Physics
• Colorado State University
• Stefan Vasile
• aPeak
• 63 Albert Road, Newton, MA  02466-1303

2
Motivation

• Scintillating fiber, or WLS readout of
  scintillator strips basic component of several
  existing detectors (MINOS, CMS-HCAL) option for
  LCD
• Standard photodetector photomultiplier tubes,
  great devices but
• Expensive (including electronics etc.),
• Bulky, magnetic field sensitive
• Multi-anode PMTs a great step forward.
• Geiger-mode Avalanche photodiodes (GPDs)
• Pros Large pulse (volt) high quantum
  efficiency relatively fast compact low mass
  low voltage operation (10s volts) modest
  physical plant magnetic field insensitive
  compatible with CMOS -gt cheap?
• Cons High dark count rate small pixels (10-150
  microns) unproven.
• New funds available to the developer, aPeak
  (Newton, Mass.)
• Small p.o. from SLAC/CSU
• A Phase I DoE SBIR recently approved w/
  subcontract to CSU

3
Status at Arlington (01/10/03)

• Encouraging device characteristics measurements
  on small diameter GPDs (at aPeak)
• Old unpackaged 50 mm GPDs showed substantially
  increased dark count rate making them unusable
  for our studies aPeak has designed new GPDs in a
  shared IRD run using a technology less sensitive
  to moisture problems. HEP will likely use sealed
  package in any case.
• Active quenching circuitry provides 1ms - 0.1ms
  pulse widths.
• Tapered optical couplers fabricated.
• Cosmic ray test setup with MINOS-style
  scintillator/fiber from SLAC. Calibrated with 1
  pmts.
• Larger, 150 mm devices by early February, 2003.
• Applying for funds to continue the work.

4
Since Arlington

• New 50 and 150 mm GPDs received by aPeak
  February, 2003 characterized with a pulsed,
  green LED source by aPeak. Piggy-backed onto
  another customers order - only available for our
  studies for a short time.
• In March, CSU scintillator/WLScosmics test
  apparatus sent to aPeak Warner (CSU) followed
  preliminary results here.
• aPeak received SBIR Phase I award with
  sub-contract to CSU includes funds tech. support
  and for devices at CSU (finally!)

5
Device Characterization

• All characterization measurements were performed
  at aPeak. We hope to cross-check them when we get
  devices at CSU.
• Configuration for most of the results in this
  talk
• Green (550 nm) LED, 150 ns pulse_at_10 kHz
• Avg. 10 photons/pulse
• Fixed bias across GPDquench circuit, 14.5 V
• Active quench circuit (lt 1 msec output pulse
  width)
• Measurements
• Dark Count Rate, DCR
• Detection Efficiency, DE (Illuminated Rate -
  Dark Rate)/10 kHz
• Temperature dependence

6
New devices characteristics (1)
Temperature Dependence
50 mm f
150 mm f
DCR
DCR
DE
DE
DE
DCR
DCR
DE
DEDetection Efficiency, DCR Dark Count Rate
7
Interpreting Detection Efficiency
photons detected, nd QE A Ng where Ng is
the number photons incident on the photodetector
and QEA is an effective single photon detection
efficiency. From Poisson statistics, the
probability for nd to fluctuate to 0 is given by
So the we define a Detection Efficiency for a
digital device, For 150 micron GPD at 20oC,
DE0.50 for lt Ng gt10 gt QEA 0.069
8
New devices characteristics (2)
Dark Count Rate Breakdown voltage, temperature
150 mm f
50 mm f
DCR
DCR
Volts above breakdown
Volts above breakdown

• GPDs with passive quenching used for these
  measurements
• Range doesnt extend to operating region of the
  cosmic ray test configuration (solid lines
  shouldnt be used for extrapolation)

9
New devices characteristics (3)
Breakdown voltage Detection Efficiency
Breakdown voltage vs. Temp.
Detection Efficiency vs. Bias voltage
1.00
slope 13 mV/?C
11 photons/pulse average room temp.
Detection Efficiency
Breakdown, V
0.50
50 mm f
150 mm f
13.65
13.70
13.75
13.80
13.85
13.90
Temperature, ?C
Bias voltage, VR

• Breakdown voltage relatively insensitive to
  temperature
• Detection efficiency plateaus well (need to see
  behavior for higher values of VR)

10
GPD Scintillator/Fiber Test Bed
For these measurements only a single fiber was
instrumented with GPD readout.
11
GPD Scintillator/Fiber Test Bed
HV
Disc.
Ch. T
Cosmic Ray Trigger
2-Fold
(To Scalar Ch. 1 and
Coinc.
GPD Coincidence)
Disc.
Ch. B
Trigger Scintillators
2-Fold
GPD/PMT Coincidence
Coinc.
(From PMT)
Disc.
Y-11 Readout
GPD Noise Rate Monitor
Test Bar
Active
Scope
From GPD
Quench
Amp.
MINOS-style scintillator bar w/ Y11 WLS readout
courtesy of SLAC
12
(No Transcript)
13
Cosmics detection efficiency
No ADC/DAQ available at aPeak so Estimate
average number of photons/event at the end of
spliced Y11 fiber using digital scope traces from
the reference pmt (average pulse height and pulse
width into 50-ohm load). For 1 mm diameter Y11
cores and 0.15 mm GPDs
Consistent (20) with calibration at CSU with
same pmtADC/DAQ. Using QEA0.069 estimated for
150 micron GPD at 20oC, predict DE
(1-exp(-0.0694) 0.24 This neglects additional
losses, such as Fresnel reflection at the Y11-GPD
interface.
14
GPD Efficiency Procedure
We compare the triple coincidence of two
hodoscope scintillators and the GPD readout of
the test bar with that of the hodoscope alone
after correcting the triple rate for accidentals
due to the GPD dark count rate, 375 kHz.
Low statistics signal runs (4-8 hours) were taken
for three configurations with the hodoscope
positioned to provide essentially full coverage
of the test bar. These signal runs were
interspersed with background runs for which the
hodoscope was moved to give essentially zero
overlap with the test bar. These background data
were compared to the expected accidental rate
from the measured GPD dark count rate. For
configurations 1 and 2, there was a press-fit air
interface between the Y11 fiber and the GPD for
configuration 3, optical grease was used to
improve the coupling. For configurations 2 and 3,
the GPD discriminator output and trigger gate
widths were reduced, halving the probability for
accidental coincidences.
15
Very Preliminary Data
16
Comments

• These preliminary data were taken under less
  than ideal circumstances, so one should beware of
  drawing firm quantitative conclusions. With that
  caveat
• It is the first demonstration of WLS fiber
  readout with the aPeak GPDs.
• The measured detection efficiency with the Y11
  readout, 215(stat.)??(sys.), is consistent
  with that predicted using the LED measurements.
• Doubling the number of incident photons (perhaps
  with a tapered optical coupling) should roughly
  double the detection efficiency.
• Lowering the GPD operating temperature to 30ºC
  potentially will more than double the detection
  efficiency as well as reduce the dark count rate.

17
Future Plans

• aPeak funded SBIR Work plan
• Fabricate 2 runs of GPD arrays using two layout
  design concepts
• Development of active quenching circuits for
  hybrid integration
• Electro-optical evaluation of the GPD array and
  active quenching circuitry array performance,
  including reliability testing
• Continue to improve the layout design for
  increased detection efficiency, lower dark count
  rate and GPD array re-configuration on-the-fly
• Evaluation of GPD prototypes for detection
  efficiency, false counts, and timing performance
  on a cosmic ray setup at CSU
• Investigate potential for use in LCD
  Muon/Calorimeter readout.
• Review multi-pixel/fiber readout scheme(c.f. B.
  Dolgoshein, Silicon Photomultiplier)

18
aPeak Advertisement